I don't see where it suggested anyone would love it
Google says it imposes this upon its employees, and that you, as a PHB, can impose it on yours too. Doesn't suggest that they will like it (as with hot desking), just that you can make them do it.
Google thinks the time has come for widespread adoption of PCs-as-a-service, so has offered up its own experience as an exemplar how to get it done. The company’s explained that it operates a “Grab and Go program” that sees it offer racks full of Chromebooks. If a worker’s machine breaks, they just grab a new one from those …
It's a Chrome book. By definition that means no real work is done on it, just lightweight web/java app type stuff. No heavy lifting like simulation tools, etc.
It's a standard MBA-like mindset to think that your workforce can grow flowers when given shit for tools.
It's a Chrome book. By definition that means no real work is done on it, just lightweight web/java app type stuff. No heavy lifting like simulation tools, etc.
In the wide range of businesses I've worked for, the overwhelming use for computers has been "light" office productivity work that a Chromebook would be entirely adequate for. There's a tiny handful of power users who need more grunt, but these people will in any event be seeking a power notebook or a proper workstation.
From a support perspective, I suspect that a decent Chromebook would be more reliable, easier to support, more secure, and more popular with most users. For the few who want/need heavy i7 laptops, or a desktop that'll cause the streetlights to go dim, then let them have them.
"The company’s explained that it operates a “Grab and Go program” that sees it offer racks full of Chromebooks."
We have a similar concept with locked down Surface Pros / Dell XPSs, which at least can run proper local business applications, and will work without an internet connection.
"does, however, point out that Citrix and VMware are Chromebook-friendly, so can deliver such apps"
To deliver that we use Dell Wyse terminals with 2-3 screens a desk and a proper keyboard / mouse. Anyone can sit down and login with no need for a crapbook, and with a far better user experience.
"No Windows sysadmin is capable of delivering that experience because the OS does not provide for it."
Windows has provided for that for well over a decade if not two. From XP onwards it was pretty easy too. If you can't do it personally that's a training and skills issue but don't try to pretend the system can't. Windows can even push apps with policy alongside your data, so the "grab and go" experience actually allows you to leave the building and carry on working on the train. This Google version will fail as soon as you lose network connectivity.
FWIW Windows also allows users to rebuild/upgrade their device by themselves if you care to set up the infrtastructure, so when their device fails they can try that while using a spare from the shelf.
The main issue is that Windows devices are worth money so you don't generally leave a rack full lying about in the office. Chromebooks, on the other hand, are essentially worthless so no point stealing them - even the staff don't want to use them let alone selling them on!
@ac: oh, for God's sake, don't try to sell us on that wet fantasy of Windows "grab and go" because most of use aroung here have suffered from some attempt at implementing that fantasy. In the best case "grab and go" is just a reimaging of a OS installation, followed by AD policy updates, with the associated application installs. If all goes well, after a couple of hours your "grab and go" machine will be ready to work. That is, of course, if the only local app is just a Citrix client that you use to connect to a remote desktop.
Note that having a build ready in a few hours without human intervention is still way ahead of the old days. So this is not an attack on the whole concept, but really a warning: those Windows tools don't provide for the "grab and go" experience that Google describes, where a user picks a machine and is able to get back to work in a matter of minutes, not hours or days.
im a sysadmin for a large network, we have no issues here, if a computer fails, we bring along another which has our current image on it, pushed out with WDS and deployment workbench, the user logs on with their roaming profile and continues where they left off. apps and programs are pushed out via group policy.
The point is that if you can hotdesk (like it or not), with the available PC and roaming profile then you have the fundamentals of Grab and Go. And school laptops use this principle anyway. A kid gets given a laptop from the pile and logs in.
Adding connection to a remote data store ( i.e. "The Cloud") is no great earth shattering innovation. However providing a working device without access to the internet in some dead spot is a whole nother boiling vessel of aquatic sustenance.
Yep. You'll lose most (if not all) of the audience when you state "roaming profiles", because the way MS implemented it (and still has implemented it) is a bandwidth sucking dire pig trying to pull a ~200 litre drum of molasses through a very small straw.
The primary around that tar pit is VDI, in which case you are using the chromebooks as nothing more than RDP clients to a giant bestial cluster of server nodes with fusionI/O cards (or other such on-node storage accelleration) or running vSAN (or are nutanix boxen)
in which case you are still a prisoner and beholden to the Dreaded Backhoe of Doom in case the network connection gets whacked.
Yes, just exactly what sort of bloated crap does Windows store in its roaming profile?
On my Linux box, login, my filespace gets mounted from the server, job done.
On my Windows box, login, anything from 1 - 5 minutes passes while who knows what junk gets downloaded, and only then can I access my desktop (and my filespace on the server).
And ever more cruft decides to add itself to my profile, so that every so often I have to ask for it to be nuked (so it can’t have been important cruft anyway) as it apparently becomes too big to “sync” itself to the profile server. Just what is the point of all this?
"Yes, just exactly what sort of bloated crap does Windows store in its roaming profile?"
Anything ranging from app/user settings to email, files, temp files etc. I'm pretty sure you can calculate your user profile size via explorer and track what exactly is taking the space.
"On my Windows box, login, anything from 1 - 5 minutes passes while who knows what junk gets downloaded"
Ask your admin why it's taking so long. Perhaps your infrastructure has a bottleneck or the admin is incompetent?
Are your roaming profiles set to save everything or are the documents and such redirected to a server location?
Could even be some stupid misconfiguration of antivirus checking all the stuff that gets loaded and slowing everything down.
Thanks for the explanation, it still sounds idiotic, however. Rather than tediously downloading all of the crap in advance, why doesn't Windows just look for it in a known location in my home folder (which is on the network file server), and only fetch/save any of it as and when a particular file is actually needed?
It seems to be the equivalent of checking out every book in the library "just in case", instead of only taking the one that you actually want at any given moment.
(And since we don't generally use Outlook or IE, at least there's hopefully not much of that crap being stored in the profile, anyway.)
Well exactly. I forget when my team got interchangeable PCs working for the majority of users in the organisation, but it was possibly twenty years ago. Maybe more now I think of it because it was always an aim, even pre Windows. Standard PCs without specialist apps were interchangeable in the 90s. Standard desktop, network delivered apps, data on the server. We were a Novell shop so it was a dozen times easier than a pure MS network.
" the user logs on with their roaming profile and continues where they left off. apps and programs are pushed out via group policy."
That's pretty close to "grab and go", but not quite. For a start, you have your machines pre-imaged, which means someone is taking time (and money) to put your image there. Also, everyone in your environment has exactly the same client software built into the image and already installed (otherwise group policy updates will kick a series of installers) so your licensing is quite simple, and no one uses any kind of specialist software.
Your environment is likely some kind of call centre, one with very little software diversity, these kind of environment are the exception, rather than the norm.
"In the best case "grab and go" is just a reimaging of a OS installation, followed by AD policy updates, with the associated application installs"
Not if you do it right. We just install all apps on the image and hide the ones not wanted for specific users via Group Policy. It's just grab and go for anyone and about 30 seconds for first logon.
Re-imaging is a once every 6-18 months process for new Windows build releases and takes about 30 minutes a Surface Pro.
"Note that having a build ready in a few hours without human intervention is still way ahead of the old days."
I think the point is that in a large organisation, particularly admin type departments, which are often he majority, they all have the same build. We do hardware support for a number of large orgs, and that's what we see every day. Having pre-imaged hard disks for desktops (or a whole PC) and pre-imaged laptops means a swap out is an almost instantaneous fix from the users point of view leaving the actual fix to happen without a user breathing down your neck.
"I think the point is that in a large organisation, particularly admin type departments, which are often he majority, they all have the same build.".
I remember working for such a shop once. They treated the sw dev department the same. "Hey, if any can run outlook on it surely you can develop your highly specialized embedded software on it". The machine was good enough for (barely, when the anti-virus, policy enforcing crap and surveillance software wasn't maxing out the cpu) writing my resignation letter.
"so the "grab and go" experience actually allows you to leave the building and carry on working on the train. This Google version will fail as soon as you lose network connectivity."
Why would you lose network connectivity if you go on a train? This is not the 90's anymore...
Wine is what they drink, whine is what they do.
https://www.theregister.co.uk/2001/07/28/bofh_and_the_linux_evangelist/
You seriously failed to understand Simon Travaglia's pun.
"But wait a minute, you could run a Windows EMULATOR on your Linux box!! Something like Wine."
"Wine? What is it?"
"Something that users do."
"Pardon?!"
"Wine? It makes your Linux box pretend to be a Windows box again. Say, how much memory has your machine got?"
As any Unix sysadmin of old can tell you, in a correctly set up environment with $HOME on NFS, NIS/LDAP and $HOME and /usr/local/??? mapped via autofs any PC is 100% interchangeable. In the days when I ran sysadmin in a development shop we operated full grab-n-go on all Linux workstations. It took less than 3 minutes to swap one as there was nothing to do software-wise. It just worked. In fact, even that was unnecessary - people could just grab the hot-desk while their machine was being services. Windows however... you were looking at a couple of hours time for each swap.
So back to Chrome. If your data is on the network, if your authentication is from the network and you cannot swap a machine by simply logging in on a new one - you are doing it wrong. Google are demonstrating that they are doing it right. Sure, it is an achievement, but only people with windows background need to sing hallelujah. If you have run a properly setup Unix network it is a "Meh, nothing new".
I think that Google is saying they have a "properly setup Unix network" only with Linux.
The comment about Windows and XP reminds me of people coming into work, plugging in, going off to get the coffee and coming back fifteen minutes later in the hope of having a domain login prompt. In the case of one company I visited, more like 45 minutes.
That may not be the case nowadays but I think that experience put an awful lot of people off the idea of shared drives.
"The comment about Windows and XP reminds me of people coming into work, plugging in, going off to get the coffee and coming back fifteen minutes later in the hope of having a domain login prompt. In the case of one company I visited, more like 45 minutes.
That may not be the case nowadays but I think that experience put an awful lot of people off the idea of shared drives."
That sounds like the days of old when some admins didn't understand roaming profiles very well and allowed users unlimited profile space and filled up the desktop with folders full of huge files which had to be populated to the local copy of the profile instead of saving their files to the "network drive" or properly maping "My Documents" to the network storage. And huge outlook mailboxes full of PDFs and image file attachments going back years.
Minor problem here. If your network switches are MAC-address locked, pulling a PC (or Chromebook) and plugging in a new one would (in a security conscious setup anyway) would then lock out that port. So now the end user has to tell the network guys which port he's removed the device from and what the new device's MAC address is. An everyday thing for the computer department, not so for Fred in Accounts !
@Blockchain Commentard you make a good point in a legacy network. In cloud world though, ports don't need to be locked down since they only access a public network anyway. Services are secured at the service so all this cloak and dagger security becomes unnecessary. If your main security requires keeping people off of the subnet you're probably already compromised. Proper authentication, encryption etc. is more than enough for normal use-cases, and for abnormal use-cases port locking is laughably innefective so doesn't really contribute. Most devices have their MAC printed on them, and most NICs can spoof a MAC address - can you see the problem here? Even if the MAC isn't printed on, all you'd need to do would be to power up the device and plug it into your own switch - you're on the network with a spoofed MAC in seconds!
> If your network switches are MAC-address locked
... then you value obscurity over security.
If you *must* do port authentication, then use 802.1x (i.e. user has credentials to access the network)
But better to go the BeyondCorp route, and not trust the network at all. All app communication is either over HTTPS or VPN.
"If your network switches are MAC-address locked, pulling a PC (or Chromebook) and plugging in a new one would (in a security conscious setup anyway) would then lock out that port"
Places that are security conscious would generally use NAC as locking by MAC address is a close to useless approach as it takes a matter of seconds to spoof a MAC on most devices.
Yeah, though now of course the 'terminals' can be used almost anywhere. It's not that new a return, either - laptops running a locked-down Linux for accessing an organisations network (the sensitive bits) have been around for years for much the same reasons; no data is stored on a laptop that might be lost or stolen.
Sort of, except that you don't use any of the local storage at all except for user profile files that have to be on C:\. Grab a laptop, log in to domain, it downloads your user profile and voila' . In my experience the biggest problem for software in this scenario isn't the software itself but the licensing, which is still in most cases sold on a per-installed-desktop basis. So this software either has to be network-installed after user logon, or else you need to go full Citrix (but this would then negate the value of having local processing power and memory and reduce it to a true dumb terminal)
Than somebody'll reinvent the PC to be disruptive.
Indeed.
Sales and marketing guys love to have something new and awesome to sell and promote. It doesn't matter what we have now... it's never as good as what the sales guys want us to buy today. If it was, why would anyone spend all kinds of money on it? Sell, sell, sell! It doesn't matter that the new thing we're supposed to buy now is the same thing that we abandoned when whatever it is we have now came along. That was ages ago, and most people have forgotten. What's old is new again, and what used to be new has gotten old. Just give it a shiny new name, like "cloud" or "thin client", and go out there and sell it!
Yeup, every 20 years, regular as clockwork...
In the '70's we had V100's, in the '90's we had X Terminals, and now we have Chromebooks and dumb Win32 clients.
As for supposed added horse power I had a Mac in 1984 with a VT100 window sitting beside my actual VT100 where the backend VAX 11/750's were main storage. No Mac hardrive in 1984. That came in 1985.
Oh, yeah. And all PC's/terminals in the building were fully networked. Mixed bag of VT100's and graphic terminals, a whole bunch of Victor / Apricot / DEC PC's , a few Macs' and even a Lisa or two.
So what is it that's new again?